Model profile
GLM 5 Turbo
Evidence summary
Safety. GLM 5 Turbo has an estimated Safety rank of #253; its 90% source-sensitivity interval is #58–#312. Its behavior-only rank is #248; company governance moves the combined estimate to #253. Published Safety evidence spans 7 eval lineages and 6 of 7 components. Its strongest relative result is Every Model Cheats — Cybench Cheat Propensity (severe_anti_cheat_cheat_propensity_pct, #1 of 22); its weakest is Adversarial Humanities Benchmark (AHB) — Table 5 (violent_crimes, #31 of 31).
Freedom. GLM 5 Turbo has an estimated Freedom rank of #68; its 90% source-sensitivity interval is #53–#290. Published Freedom evidence spans 4 eval lineages and 1 of 1 components. Its strongest relative result is Adversarial Humanities Benchmark (AHB) — Table 5 (violent_crimes, #1 of 31); its weakest is Enkrypt AI Safety Leaderboard (harmful_attack_non_success_rate, #192 of 270).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗glm-5-turbo
- OpenRouter ↗z-ai/glm-5-turbo
- Official model documentation ↗Family-level model document · zai · first party
- Release source ↗curated outward evidence
Safety evals
Rank and direction are specific to the Safety portfolio. Black marks the observed result; the grey dot marks the value implied by the Safety rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Source |
|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #113 / 358 | ↓0.6264 | Source ↗official | |
| Adversarial Humanities Benchmark (AHB) — Table 5child_sexual_exploitation | #16 / 31 | ↓48.9 | Source ↗official | |
| Adversarial Humanities Benchmark (AHB) — Table 5defamation | #30 / 31 | ↓88.6 | Source ↗official | |
| Adversarial Humanities Benchmark (AHB) — Table 5hate | #27 / 31 | ↓71.7 | Source ↗official | |
| Adversarial Humanities Benchmark (AHB) — Table 5indiscriminate_weapons_cbrne | #26 / 31 | ↓86 | Source ↗official | |
| Adversarial Humanities Benchmark (AHB) — Table 5intellectual_property | #29 / 31 | ↓80 | Source ↗official | |
| Adversarial Humanities Benchmark (AHB) — Table 5non_violent_crimes | #29 / 31 | ↓93.5 | Source ↗official | |
| Adversarial Humanities Benchmark (AHB) — Table 5privacy | #26 / 31 | ↓80 | Source ↗official | |
| Adversarial Humanities Benchmark (AHB) — Table 5sex_related_crimes | #19 / 31 | ↓73.9 | Source ↗official | |
| Adversarial Humanities Benchmark (AHB) — Table 5sexual_content | #21 / 31 | ↓55.6 | Source ↗official | |
| Adversarial Humanities Benchmark (AHB) — Table 5specialized_advice | #26 / 31 | ↓73.2 | Source ↗official | |
| Adversarial Humanities Benchmark (AHB) — Table 5suicide_self_harm | #26 / 31 | ↓69.6 | Source ↗official | |
| Adversarial Humanities Benchmark (AHB) — Table 5violent_crimes | #31 / 31 | ↓93 | Source ↗official | |
| BullshitBench v2clear_pushback_rate | #71 / 122 | ↑0.27 | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #214 / 270 | ↑11.89 | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #214 / 270 | ↑76 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #78 / 270 | ↑88.89 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #163 / 268 | ↑95.45 | Source ↗official | |
| Every Model Cheats — Cybench Cheat Propensitybaseline_cheat_propensity_pct | #9 / 22 | ↓26.1 | Source ↗official | |
| Every Model Cheats — Cybench Cheat Propensitysevere_anti_cheat_cheat_propensity_pct | #1 / 22 | ↓0 | Source ↗official | |
| Every Model Cheats — Cybench Cheat Propensitystandard_anti_cheat_cheat_propensity_pct | #3 / 22 | ↓4.3 | Source ↗official | |
| SM-Benchadversarial | #75 / 92 | ↑77.56 | Source ↗official | |
| SM-Benchambiguous_interpretation | #92 / 92 | ↑13.69 | Source ↗official | |
| SM-Benchanti_hallucination | #92 / 92 | ↑18.85 | Source ↗official | |
| SM-Bencheq_boundaries | #24 / 92 | ↑69.1 | Source ↗official | |
| SM-Benchoverfit | #54 / 92 | ↑70.49 | Source ↗official | |
| WildClawBench Safety & Alignment (OpenClaw harness)safety_alignment_score_pct | #21 / 24 | ↑32.19 | Source ↗official |
Freedom evals
Rank and direction are specific to the Freedom portfolio. Black marks the observed result; the grey dot marks the value implied by the Freedom rank. Values stay on each source’s native scale.